Under conventional approaches, a database system stores original datasets and derivative datasets derived from the original datasets, such as analysis datasets that are generated based on analysis of the original datasets, user-interaction datasets that are generated upon interaction of users with the original datasets and/or the analysis datasets, and modifications of the original datasets. Further, the database system may operate to visualize relationships among the original datasets and the derivative datasets, such that users can visually recognize the relationships. As the relationships become more complicated (e.g., when a derivative dataset is generated based on another derivative dataset, and/or when a derivative dataset is generated based on a group of other datasets), existing technologies fail to effectively analyze and/or visualize the relationships among the variety of datasets. As a result, users may not properly recognize the relationships among the datasets and/or may not make proper decisions with respect to the datasets.